Some months ago I filed an application for a national patent (in an EU country). The examiner pointed out several amendments to be made, which I did after I discussed with him twice.

Eventually the application was rejected at the examination stage (i.e. before reaching the prior art search) on the grounds of exclusions from patentability. Afaik there is no way to appeal.

How would it be best for me to proceed?


Relevant details:

The rejection invoked the exclusion area of scientific theories and software. The material is actually neither, although it could be mistaken for them by someone not familiar with the field.

One of the pillars in the (attempted) patent is the application of machine learning methods, which the examiner sounded like he isn't quite familiar with. At one point during our discussions he exclaimed "do you mean to tell me that you'll input data and you'll just get an answer?", which is exactly what machine learning is supposed to be about when handled properly.

As an aside, this is not about improving a machine learning algorithm or about a trivial application; unfortunately you'll have to take my word for it for the time being, given the situation :)

An acquaintance, an engineer who collaborates with patent attorneys, read the application and agreed that there is no ground for rejection. He added that it's the first time he hears about an application being rejected before the prior art search.

All in all, I believe that the rejection is due to a combination of novelty, unfamiliarity of the examiner with the field, and rather abstract nature of the material (although its practical applications in industry were described in detail).

Lastly, during the examination period I contributed a poster to a conference, which included a description of the main concept used in the patent.


At the moment I'm speaking with patent attorneys in order to find the best way forward, but I'd like to hear the opinion of people with similar experience or expertise...

  • 1
    If you have some time, read the epo enlarged board of appeals decision g3/08 (the second half, not the international treaty stuff). Rejecting an application on the basis of not being technical is ... ridiculous tbh. The current standard is ignoring all non technical features for inventive step considerations. But without reading the exact stuff Id be just guessing.
    – user18033
    Apr 12, 2019 at 16:45
  • @DonQuiKong , thank you! indeed a European patent is one of the ideas on the table; it'll be good to get more educated about it.
    – Helen
    Apr 13, 2019 at 18:04
  • 1
    Oh I misread the eu country part. In that case it depends on the jurisdiction. Germany is similar to the epo on that, the others I don't know. But it really doesn't matter if you get caught on technicity or inventive step, you need to show that the relevant features are technical.
    – user18033
    Apr 13, 2019 at 20:56
  • @DonQuiKong I'm sorry but I miss your last sentence - would you mind explaining it more? "You need to show ... are technical" reads to me like the opposite of "rejecting on the basis of not technical is ridiculous".
    – Helen
    Apr 15, 2019 at 15:34
  • 1
    It is. Well, kinda. At the epo an invention is technical if any feature is technical (so “computerimplemented“ is enough). BUT only technical features are used for inventive step. Means: a new computer game is technical, but not inventive because the game itself isn't technical. But that's an inventive step rejection, not technicity.
    – user18033
    Apr 16, 2019 at 18:01

2 Answers 2


I'll take a stab at answering this despite the vague language in your question, and the laws coming from an anonymous EU jurisdiction.

The EU under the patent articles does in fact forbid software programs, "The following in particular shall not be regarded as inventions ... [including] ... programs for computers." Art. 52(2) EPC.

You don't indicate how you're claiming the subject matter or the scope of your patent application.

The suggested route for your rejection would be either:

  • amend the claims so that they are not "programs for computers" per se, but rather are directed to an apparatus or a "computer-implemented method" OR
  • IF your application does not provide support for an apparatus or computer-implemented method, you should file a new application including such material that claims priority to the parent application.

"An acquaintance, an engineer who collaborates with patent attorneys, read the application and agreed that there is no ground for rejection."

Sorry but these statements always amuse me. Your acquaintance isn't a patent attorney/agent, much less one familiar with the local rules of the jurisdiction. His assurance that there's no ground for rejection is as useful as the answer from an 8-ball.

edit: Want to add some additional context for machine learning (and AI) type patent applications. The recent Guidelines for Examination in the European Patent Office (November 2018 edition) included a section on AI and machine learning. (See Section 3.3.1 Artificial intelligence and machine learning)

Key quote:

Artificial intelligence and machine learning are based on computational models and algorithms... [s]uch computational models and algorithms are per se of an abstract mathematical nature...

In other words, your machine learning application will be rejected under Art. 52(2) EPC, unless you have some "technical application" or "technical implementation" of the machine learning methods.


  • 1
    email to [email protected] bounced back
    – George White
    Apr 20, 2019 at 17:29
  • Wow, thanks for the information. Will have it fixed shortly. In the meantime, [email protected] is my personal inbox.
    – Chris
    Apr 21, 2019 at 0:07
  • Hi Chris, first of all thanks for giving an answer. However I should clarify some points (and disagree, I'm afraid, with a few ones by you in the process). If you read the first remark under "relevant details" you'll see that this is not about a program for computers (okay, I called it software; I might not be familiar with English legal terminology, but I can understand the context and meaning of my application ;). Second, why did you raise the issue of "laws coming from an anonymous EU jurisdiction"? If you read the question again, I don't plan to appeal (mainly because I don't have the
    – Helen
    Apr 22, 2019 at 17:33
  • right to, and this I know for a fact). Apologies, but I'd rather keep any detail that is not tremendously relevant out of this forum for now (i.e. the country). This includes the subject matter and the scope. Third, I'm really rather taken aback by your last paragraph; I'm not asking for legal advice here and I'm not looking for appearing smart while throwing out irrelevant info. I'm sorry to say that your remark is a little... well, negative. Also, that guy is perfectly capable of discerning whether my application pertains to a scientific
    – Helen
    Apr 22, 2019 at 17:40
  • theory or a program. But again, this was not my point, it was just extra added background to my question. I see that you are rather new in SE and I hope that you provide more relevant remarks in other answers.
    – Helen
    Apr 22, 2019 at 17:42

Each country has its own patent rules. What is allowable in one country may not be allowable in another country. There may be number of reasons why it is rejected. For example in some countries, there is formality check (i.e. does the application corresponds to the formal requirements) before the prior art search. Best thing you can do is to consult with local patent attorney.

  • Thank you for the reply, but actually my question is about how to proceed best now. (I.e. I am not asking about justification or not of the decision, or appealing the decision.) Just for the sake of example, suggestions could include: reapply with certain changes, apply for a more international patent, etc. (PS: As I mention in the question, I am already consulting with patent attorneys, but I'm also interested in SE's opinions.)
    – Helen
    Apr 13, 2019 at 17:59

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .